1The Argument in a Nutshell

DEMOCRACIES almost never fight wars against each other. This simple observa-tion is one of the most powerful findings in international politics and one ofthe most throughly tested. But what explains it? The answer, I think, is thatdemocracies have unique “contracting advantages,” which allow them to buildstable, peaceful relations, based on multiple self-enforcing bargains.

The basic finding—that democracies rarely, if ever, fight each other—is oneof the most compelling in modern international politics. Discovered by acci-dent and initially overlooked, the “democratic peace” has been vigorously de-bated and exhaustively tested. Most statistical studies and case histories havefound the same robust relationship. Democracies often go to war but veryseldom against each other.1

Thanks to all this testing, the democratic peace is now one of the best-established regularities in international politics, perhaps the best-established.2

The absence of war between democracies, Jack Levy concludes, “comes asclose as anything we have to an empirical law in international relations.”3

Another recent article notes that when all interacting states are democracies,that is “a near-perfect sufficient condition for peace.”4

Not surprisingly, these findings have prompted a flurry of research. Someextend the initial conclusions, showing that democracies are distinctive in otherways. They generally win wars, for example, probably because they take spe-cial care in choosing opponents.5 Others show the limits of the democraticpeace. New and unstable democracies may well be more war-prone than otherstates, although they, too, are reluctant to fight among themselves.6 This doesnot mean that democracies are pacifists, however, even when dealing with eachother. They sometimes threaten to use force against fellow democracies, andthey have come close to war several times.

Even with these qualifications, the democratic peace is a powerful findingwith far-reaching implications for both policy and theory. It means that interna-tional interactions are profoundly shaped by the way states are governed. Natu-rally, this finding inspires democracy’s many advocates. But it deeply puzzlestheorists, who want to know why it occurs. Their puzzlement echoes an oldacademic joke: “We know it works in practice. Now we have to see if it worksin theory!’ ”

That is exactly the question about peace among democratic states. It workswell in practice, but there is considerable confusion about how it works in

C H A P T E R 12

theory. The lack of an answer is no joke, however. Despite extensive research,all we have is a remarkable correlation. We still lack a convincing explanationabout why democracies do not fight each other.

Knowing why they do not fight is important for both practitioners and theo-rists. It bears directly on two central issues of international politics: the reasonsfor war and peace and the problems of sustaining cooperation. It also hasserious implications for any general theory of international relations. It posesa specially pointed challenge to those who deliberately ignore the character ofdomestic governments, a crucial simplifying assumption of neo-Realism.7 Theanswer to this puzzle may also say something important about how countriescan build relationships of greater trust and reassurance, laying the foundationsfor enduring peace. In short, it is a prominent puzzle in every sense.

How do we explain this apparent relationship between governmental formsand international outcomes? So far, three basic explanations have been ad-vanced:

1. citizens’ reluctance to bear the costs of war;2. shared values among democracies; and3. unique domestic institutions, which restrain elected leaders.

The cost explanation, initially developed by Immanuel Kant, argues thatcitizens of a republic are less war-prone because they must bear the burdensthemselves and can vote to avoid them. Monarchs and dictators, by contrast,can shift the costs of war onto their subjects, who have no voice in thedecision.8

The normative explanation, also suggested by Kant, is that liberal democra-cies share certain basic values, grounded in their domestic political life. Theysettle disputes through neutral courts rather than through blunt force or statusdifferences.9 They place a high value on individual life and liberty. As theylook abroad, they recognize that other democracies also have governmentsbased on popular consent and hold similar values. They then apply these liberalvalues in dealings with other democracies, where they can expect reciprocity.10

Relying on these common values, they can adjudicate disputes and compro-mise voluntarily rather than resorting to military force.11

Finally, the institutional explanation underscores the constraints facing dem-ocratic policymakers.12 They face constitutional limits, must share power withother elected leaders, and can remain in office only by winning periodic elec-tions, openly contested. Although there are important differences between par-liamentary and presidential systems, both include legitimate opposition, genu-ine public debate, and procedural limitations on how public policy is made.13

These institutional arrangements inevitably slow down decisionmaking andmake it difficult to launch military operations. If other states are similarlyconstrained, then neither will fear surprise attacks, and both will have the lux-ury of time to resolve international crises.14

T H E A R G U M E N T I N A N U T S H E L L 3

All three explanations highlight important features of democracy, but eachis incomplete. If costs alone were the answer, then citizens would be equallyreluctant to fight both democracies and nondemocracies. That does not explainwhy they choose to fight some types of states and not others. Large, powerfuldemocracies have waged wars against small dictatorships, but not against smalldemocracies. Moreover, their sensitivity to costs is not so straightforward. Pub-lic opinion in democracies sometimes favors war, as it did in 1898, when theUnited States fought Spain, which was blamed for sinking the USS Maine inHavana harbor. Monarchs and dictators, on the other hand, are not alwayseager to fight, since they, too, must weigh the financial costs, political risks,and chances of success. Defeat will not hurt them at the polls, but it could leadto their overthrow, exile, or death.15

If norms alone were the answer, democracies would also be more peacefultoward all other states, not just other democracies. There is now some evidenceto support that assertion, at least in mild form. Democracies are probably a bitless war-prone than other states. Even so, it is clear that democracies go to waroften. In the twentieth century, for instance, Britain went to war more fre-quently than Germany.16 The United States has obviously been willing to useforce repeatedly, both to defend its material interests and to extend its values.Moreover, democracies do not merely respond to provocations by others. Theyoften seize the initiative and sometimes attack first. We need to understand,then, how norms and beliefs could lead democracies to make war against somestates but not against others. If their norms encourage them to negotiate, com-promise, and use judicial mechanisms, why don’t they do exactly the samething with nondemocracies, achieving peace with them, too?

Equally important, if democracies are really bound by such strong, pacificnorms of conduct toward each other, why do they sometimes threaten fellowdemocracies with force? Contrary to the normative explanation, democraciessometimes issue military threats against each other and display troops andweapons to make them more credible. They probably do so less often thannondemocracies, but they definitely do issue military threats against each other.They have occasionally launched covert military operations against other de-mocracies or supported coups from within. A purely normative explanationcannot easily grasp this kind of belligerence or fold it into an account of thedemocratic peace. What we need is a much better understanding of how norma-tive concerns operate, how they dovetail with material interests to define na-tional preferences, and, ultimately, how they shape the choices that yield warand peace.

In fact, common norms, shared values, and even shared interests may not besufficient to produce peace. They may foster cooperation, but they do not en-sure it. That is the fundamental lesson of the Prisoners’ Dilemma. Both playersvalue mutual cooperation over mutual defection, but they still cannot achieveit in a finite game. To achieve stable cooperation, the game itself must be

C H A P T E R 14

changed in fundamental ways. Assuming both players value the future highlyenough, they can sustain cooperation only if the game is played repeatedly,without a clear end point, and if defection by either player is deterred by thethreat of punishment (that is, by the threat of noncooperation later in the game).

Finally, if domestic institutions were a sufficient explanation, why aren’tdemocracies constrained from using force against all types of regimes? Notonly do they use force against nondemocracies, they sometimes strike the firstblow. The larger problem here is that institutional explanations do not spellout the causal connection between domestic limitations and international out-comes. How exactly are constraining institutions supposed to prevent democ-racies from going to war? And why do they operate to prevent wars only withother democracies?

All three explanations—costs, norms, and institutions—highlight unique el-ements of modern democracy and provide insights into the democratic peace.They suffer, however, from the same basic problem. Each is focused on theproperties of an individual democracy. The democratic peace is fundamentallyan interactive phenomenon. It is not about why one democracy or another ispeaceful. It is about why two democracies so seldom fight each other.

The existing explanations merely allude to this central, interactive elementof the democratic peace without developing and pursuing its implications. Es-sentially, they say, “if state A behaves this way, or holds these values, or hasthese institutions, and if state B does too, then we will have peace.” Sometimesthese explanations go one step further and say that states A and B must recog-nize their shared values or constraints. That is correct, as far as it goes, but itdoes not go far enough. It glides over the most distinctive feature of the demo-cratic peace: its emergence from the durable bargains and mutually profitablerelationships democracies have forged with each other. That is the heart of myown explanation.

My explanation is that constitutional democracies have a special capacityto make and sustain promises with each other, including those about war andpeace. They are better equipped to find and capture gains from mutual interests,to sustain them, and to forge durable, mutually profitable relationships. Demo-cratic norms and institutions facilitate this by making bargains easier to iden-tify, less risky, and more reliable in practice.

Because democracies have unique “contracting advantages,” they can usu-ally avert or settle conflicts with each other by reliable, forward-lookingagreements that minimize the dead-weight costs of direct military engage-ment. To do that, states must be confident their partners will live up to theirpromises or, if they do not, that they can protect themselves from the risks.

After all, in matters of war and peace, there are no effective global courts topunish violators or protect innocent parties. It is a self-help system. States mustlook first to themselves for protection in a dangerous world. That usually

T H E A R G U M E N T I N A N U T S H E L L 5

means arms and fortifications. But states can do more than deter, defend, andattack. They can also help themselves by making mutually profitable dealswith reliable partners, both as military allies and as trading partners.

This contracting explanation recognizes that democracies have mixed inter-ests, sometimes overlapping, sometimes diverging. They have good reasons topush hard for their own favored outcomes and may bluff and threaten to achievethem. On the other hand, they also have strong incentives to capture joint inter-ests and prevent the breakdown of cooperation. One of the most important ofthese joint interests is avoiding the barren costs of war. Their ability to achievesuch cooperative outcomes by peaceful negotiations should increase over timeas partners learn more about each other and develop profitable working rela-tionships. As they learn, their contracting relationships evolve.

This explanation does not discard the extensive research on democraticnorms and institutions. Quite the contrary; it systematically incorporates thatwork as part of a more comprehensive contracting explanation.

My explanation begins by noting that states must negotiate and secure theirown bargains, ranging from formal treaties to informal working relationships.To do that, they must locate some common interests, if any are to be found,and divide the potential gains. They must also take care to ensure that theirdeals (and, indeed, their wider relationships) are self-enforcing. Self-enforce-ment is vital because there are seldom neutral third parties to interpret andenforce international bargains. They must seal their bargains in the harsh worldof international anarchy, where states must help themselves, interpret their ownbargains, and enforce whatever deals they make.

In this setting, perceptions and beliefs matter because states are not com-pletely informed about each others’ preferences and behavior, much less aboutthe future contingencies they will face. They must also be concerned abouttheir partners’ institutional capacity to make commitments today and fulfillthem tomorrow. The danger is that partners are unable or unwilling to live upto their promises, perhaps because the leader who made the agreement changeshis mind, perhaps because circumstances have changed, or perhaps becausethere is a new ruler or a fundamental change of government.

These concerns arise because in the real world bargains are necessarily “in-complete contracts.” No matter how elaborate treaties are, they cannot antici-pate all future possibilities, all conceivable changes or external shocks. Negoti-ators either fail to foresee them or deliberately choose not to spend scarce timeand money dealing with remote eventualities. If these contingencies arise, orif the parties later interpret their bargain differently, then they have to sort outtheir disputes, adjust the terms, settle on a common interpretation, or, failingthat, abandon the agreement. When disputes arise, then, they must look afterthemselves. If they wish to capture the gains of cooperation and facilitate theirongoing relationship, they have to create some kind of governance mecha-nisms—formal or informal—to cope with such problems.17

C H A P T E R 16

Before relying on such incomplete bargains, prudent states must make judg-ments about their partners’ integrity and reliability. They want to reap thepromised benefits, but they also want to be sure others are dealing in goodfaith and will continue to do so after the bargain is struck. Is their potentialpartner stable and dependable? Can its performance be monitored effectivelyat a reasonable cost? Is it able to fulfill its commitments, and will it try? Is itlikely to discard its obligations if they become inconvenient, or cheat if itsactions are hidden or too costly to punish? When new leaders come to power,will they be bound by earlier promises? Can the participants forge governancearrangements to sort out disputes and fill in gaps when they arise? These arehard questions with serious consequences. For all these reasons and more,states are concerned about the types of partners they are dealing with. That isparticularly true in sensitive national security issues, where cheating and be-trayal can be a matter of life or death.18 That is why states look to their partners’reputations for compliance or opportunism, their transparency to outside ob-servers, their continuity of government, and their institutional capacity to makesolid commitments about future behavior.

There are always risks, of course. But they are lower and more manageablewhen partners are well-established constitutional democracies. Such states aresimply better equipped to make long-term commitments, to sustain them acrosschanges of political leadership, and to show others their true preferences andtheir compliance with existing bargains.

What drives the democratic peace, then, is not so much democracies’ dis-tinctive policy goals—although they may well be less greedy, aggressive, andconflictual—but rather their special ability to locate and capture joint gainsand to avoid deadly escalation in dealing with each other.

The key to this explanation is the reciprocal exchange of promises. To work,these promises must be credible, reliable, and durable. Otherwise they wouldnot reassure anxious neighbors. Stable, constitutional democracies have a greaterinstitutional capacity to make such reassuring, self-binding promises, both for-mally and informally. Their constitutions are designed to make some policiesvery difficult to reverse, sometimes by requiring supermajorities to change keydecisions. They also make major strategic choices visible to all, including othercountries, largely as a by-product of public discussion among elected officialsand their communication with voters. Because political opponents can openlychallenge government policies, other states can judge the depth of support forthese commitments and the likelihood, if any, that they will be overturned. Thatgives partners some protection against devastating surprises.

The mass electorate supports these arrangements in several ways. First,elected leaders generally pursue voters in the center of the political spectrum.This concern with the “median voter” serves as an anchor against dramaticpolicy changes despite turnover in elected leadership. Second, to win votes,politicians must articulate major policies and commitments to the public,

T H E A R G U M E N T I N A N U T S H E L L 7

which spreads information to friend and foe alike. Third, promises made tothe electorate serve as a kind of surety bond. Politicians know it is costly tobreak major promises, including treaty obligations. That allows them to makesome foreign commitments more credible and trustworthy by deliberately cre-ating high “audience costs” among the electorate and often among legislatorsas well. These audience costs mean that democratic leaders have political in-centives to fulfill their commitments, and their partners know it. These incen-tives, plus the transparency of democratic government, lessens the chancesthey are bluffing or deceiving.

Even so, democracies do change course and occasionally break promises.When they do, however, the process is typically slow and open to outsidescrutiny, which diminishes the risks to partners. The process is slow becausedemocracies have more elaborate, codified procedures for policymaking andmore locations where policy can be vetoed or amended. The process is moreopen because democracies are organized to communicate policy alternativesto their own citizens and their multiple representatives.

Democracy is a vast, information-generating machine, much like the marketas Friedrich von Hayek describes it. Engaging the public in major policy de-bates is a fundamental part of the electoral process, which is why press freedomis essential in constitutional democracies (and a danger to autocracies). Legis-latures also require extensive information to perform their duties. Their debatesand oversight of the executive continuously disclose information to the publicand, as an inevitable by-product, to foreign friends and foes.

Opposition parties promote this flow of information in several ways. Theyprod the government to release data, forecasts, and evaluations. They debatethe wisdom of specific policies and broad strategies. They support or opposegovernment policies and, in doing so, signal the depth of national support forthem. Again, all this information is freely available to both allies and adversar-ies.19

The resulting transparency means foreign partners can gauge the depth of ademocracy’s commitment to specific policies, not only today but over time, aspolicies are implemented.

This combination of open debate and slow policymaking gives partnersample notice of potential changes. That lowers the risks of devastating surprise,which makes it more feasible to make bargains in the first place. As the come-dian Henny Youngman once said, “It’s good to see you. It means you’re notbehind my back.” That is exactly the point about transparency, and that is whyit is so valuable in making secure international bargains.

These institutional arrangements make it hard for democracies to act nimbly,but they also make it easier for democracies to contain risks and make durablecommitments. When democracies work with partners that can reciprocate, theycan profitably exchange promises of peace—promises they fully expect to bemutually rewarding and self-enforcing for the long haul.

C H A P T E R 18

Extensions: Implications of a Contracting Theoryof the Democratic Peace

If this explanation is correct, it should not only clarify why democracies sorarely fight each other, it should tell us much more about democracies in worldpolitics. We should expect a series of other major empirical findings, closelyrelated to the contracting explanation. These extensions are important for bothsubstantive and methodological reasons. First, they indicate that the demo-cratic peace has wider ramifications for international politics. Second, theyallow us to evaluate the contracting explanation beyond simply retrofitting itto familiar evidence. These extensions are crucial to what Imre Lakatos callsa progressive research program.20 Instead of simply tailoring a theory to fitwell-known facts, he urges us to see if the theory also yields new and unex-pected discoveries. That is a hard test, but meeting it gives us confidence thatthe theory has real explanatory power.

If secure contracting really does explain the democratic peace, then weshould also expect to find the following:

• Bargaining among democracies should show less concern with sudden be-trayal or opportunistic abandonment than similar bargaining involving one ormore nondemocratic states. Negotiations among democracies should pay lessattention to mechanisms to guard against cheating, betrayal, and renunciation.21

• Bargaining among democracies should be less concerned with problems ofcompliance, verification, and safeguards (although they are unlikely to ignorethese matters entirely). These issues should be much more prominent whennondemocracies are involved. When security bargains are reached with non-democratic states, which are more opaque and less trustworthy partners, theyshould include much more extensive mechanisms for self-protection.Agreements with them must be amply buttressed against potential breach, ifthey are to be relied on.• Although the risks of opportunism and nonperformance are lower amongdemocracies, they are not entirely absent. Since democracies lack perfect infor-mation about each other and are not perfectly trustworthy, bargains amongthem should include some safeguards. Nevertheless, these measures should beless extensive, less costly, and less vital to sealing the bargain than thoseneeded when nondemocracies are involved.• Alliances among democracies—a crucial type of security agreement—shouldbe more stable and reliable than alliances involving nondemocratic states.• Changes of political leadership should have a greater impact on bargainsinvolving nondemocracies. Leadership changes should have less effect on bar-gains and relationships among democracies since new policymakers are moreconstrained by prior commitments and constitutional rules (including votingrules that heighten the influence of median voters and therefore stabilize policychoices).

T H E A R G U M E N T I N A N U T S H E L L 9

• Democratic bargains should be most reliable when they involve❍ Two states dealing directly with each other about❍ Bilateral issues that the two parties themselves can control.

They should least reliable when multiple partners are involved and whenissues lie beyond their immediate control. Agreements involving multiple part-ners pose problems because all states, including democracies, have incentivesto free ride, to let others shoulder the burdens of collective action. Democra-cies, like other states, have incentives to shirk individual expenses, pass thebuck to others, and avoid the costly task of punishing those who violate theircommitments.• The democratic peace should have the weakest hold over (a) new and unsta-ble democracies and (b) marginal democracies. Both are less reliable partners.Even if they share the normative commitments, transparency, and formal insti-tutions of older democracies, their preferences and future behavior are stillhighly uncertain. Reliance on them is risky. Potential partners cannot be sureif the domestic regime will survive, if it will remain democratic, or if it willbe replaced by a radically different alternative. As Vice President Dan Quayleonce remarked, “We are on an irreversible trend toward more freedom anddemocracy, but that could change.” It certainly can change for individualstates. Aside from the risk of rebellion and coups d’etat, new democracies areless reliable because they have no track record or long-term relationships tobuild on. This same shroud of uncertainty surrounds states that are only mar-ginally democratic, especially if partners are unsure about what type of govern-ment they will have in the future. They, too, should be largely outside thedemocratic peace because of their instability and unreliability as partners.• As a closely related proposition: the democratic peace should be weakerwhen democracies are internally threatened and unstable, when they must con-front enormous economic and political shocks. During such periods (or inunstable regions), greater uncertainty about the permanence of democratic re-gimes weakens their reliability as long-term partners.• Among well-established democracies, peace should become increasingly sta-ble over time as they develop confidence in their relationships as contractingpartners. This learning effect is an integral feature of contracting. It develops asstates gain experience in their relationships and gain knowledge about their part-ners. They learn from their successes (or failures) in resolving conflicts short ofwar. They gradually acquire more information about their partners’ preferencesand capabilities. As they construct a nexus of successful bargains and sustainthem over time, they also build confidence in each other’s trustworthiness.

This evolution resembles bargaining and cooperation among firms. “Theliterature is agreed that interpersonal trust in business relations is rarely offeredspontaneously but requires an extended period of experience,” according toindustrial sociologist Christel Lane. “During this time, knowledge about theexchange partner is accumulated through direct contact or is acquired indi-

C H A P T E R 110

rectly through reliable third parties.”22 We expect the same kind of learningamong states. Partners that resolve a series of major problems successfully,without fighting, should gradually develop deeper, more stable forms of coop-eration. Thus, lengthy periods of joint democracy should produce increasingpeacefulness, rationally grounded in experience. Even though states may stillhave significant policy differences—differences that could grow as well asdiminish—military threats should decrease markedly within such long-stand-ing democratic partnerships. This learning effect predicts, for example, that theunbroken relationship between the United States and United Kingdom shouldproduce a stronger, deeper peace in the twentieth century than in the ninteenth.• Bargains and long-term relations among democracies should not be free ofdiplomatic rivalry or strife. After all, these are independent states, and theyhave a mixture of common and conflicting interests. These mixed interestshave important consequences in the contracting explanation of the democraticpeace. States want to capture joint gains through cooperation, but they alsowant to win the lion’s share of any gains. To prevail in these distributionalstruggles, democracies may signal their resolve by issuing threats, even againstfellow democracies. Likewise, they may use threats to ensure compliance withexisting bargains. Our contracting explanation does not preclude militarythreats or even the occasional use of force in relations among democracies,but it does expect them to be very rare.

Democracies should be especially reluctant to threaten military actionagainst each other for at least two reasons. First, democracies are more trans-parent, so there is simply less private information to reveal. Second, whendemocracies need to show their private level of commitment, they have morepolicy instruments to do so. Because their societies, polities, and economiesare more open, they have more extensive transnational ties with each other.23

That allows them to use a wide range of sanctions to show their resolve.24 Itgives them valuable options short of military action, which carries high costsand risks.• Because established constitutional democracies are better able to governtheir relationships and settle conflicts by reliable agreements, they should per-form better toward each other than do nondemocracies at every level of inter-state conflict. Relationships among democracies should involve

❍ fewer military threats,❍ fewer actual displays of force,❍ fewer skirmishes or low-level conflicts, and❍ fewer outright wars,

compared to relationships involving nondemocracies.• The contracting explanation is fully consistent with defensive anxiety aboutnational security, even in relationships among democracies. Yet these Realpoli-tik fears should be dampened among democracies for two main reasons.

T H E A R G U M E N T I N A N U T S H E L L 11

❍ First, the inherent openness of democracies allows outsiders to observetheir foreign-policy choices and military posture. Their openness shouldreduce “worst-case” fears and lessen any exaggerated concerns about theiraggressive motivations and expansive (offensive) policies. That is, it shouldfacilitate more realistic and informed assessments.

❍ Second, as positive contracting experience builds over time, partners’fears should recede, assuming the experience is pertinent.

Both effects should reduce the security dilemma for any partner of a demo-cratic state. The effects should be greater, however, in relations with otherdemocracies, where mutual transparency and diplomatic experience shoulddiminish suspicions and encourage more accurate inferences. Among demo-cratic partners, these effects short-circuit the cycle of fear and response thatso often characterizes international politics.• If democracies do go to war against each other, they are much more likelyto fight about fundamental, existential issues, which are extremely difficult toresolve, rather than about peripheral matters. They are unlikely to fight eachother because of confused perceptions or inflated fears.

❍ Democracies should be better at locating profitable compromises,where they exist. Such compromises should be more accessible for periph-eral matters. As a result, costly conflicts over such matters should be rare.

❍ Because democracies have better information about each other and lessprivate information about their own preferences and intentions, they shouldbe less likely to fight each other because of confusion and misinformation.

Taken together, these mean that if democracies do fight each other, their dis-putes should deal with basic issues of physical and territorial security and corepolitical issues, central to the organization of state and society, not secondaryissues like colonial expansion.

These inferences cover a wide range of international relationships. If theyare borne out, they would give us more confidence in the contracting explana-tion. They would meet Lakatos’s test of uncovering new and unexpected facts.They would also indicate our theory has real breadth, a useful test proposedby Mancur Olson. Theories are much more persuasive, he says, when they canexplain a diverse array of facts.25

Four Sources of Democracies’ Contracting Advantage

All these inferences—which we will test later—assume that democracieshave a systematic advantage in locating and sustaining international bargainswith each other. This contracting advantage is no fluke or historical accident.It is deeply embedded in the character of democratic states and their setting

C H A P T E R 112

within civil society. Four basic traits are most important for securing interna-tional agreements:

1. democracies’ openness to outside scrutiny;2. the continuity of democratic regimes;3. visible electoral incentives for leaders to keep important promises; and4. the constitutional capacity to make enduring commitments, which can-

not be easily overturned by current elected leaders or their successors.

These are deep-seated features of established constitutional democracy, fea-tures that facilitate international agreements and lessen fears. They allow de-mocracies to reassure each other in ways that nondemocracies simply cannot,creating an informal governance structure for their bilateral relationships. Inother words, the institutional capacities of constitutional democracies allowthem to strike mutually profitable bargains, erect mechanisms for durable co-operation, and resolve conflicts peacefully.

While no international promises are ironclad, those of democracies are morereliable because they come from institutions that transcend individual leaders,are likely to persist, and are relatively open to outside inspection. The resultis that democratic promises are generally more convincing and enduring. Theyare less subject to surprise breach.

This does not eliminate the pulling and hauling of international politics, thedivergence of interests, or even the occasional threat of force among democra-cies. Diplomacy among democracies is not a Quaker Meeting. But it is not abrutal Hobbesian anarchy either, where lethal danger lurks behind every tree,promises are worthless, and treaties are mere scraps of paper. Rather, democra-cies have a unique capacity to exchange promises based on strong foundationsof rational confidence and lower risk, relying on credible commitments, notmere hopes and dreams.26

It is this foundation that allows democracies to move beyond the immediatetask of safeguarding individual promises and short-term relationships. As de-mocracies become more confident in their mutual dealings, they reap a majorbenefit: they can promote wider, more enduring ties of interdependence atlower risk. Their dealings give them considerable information about each oth-ers’ political and economic interests and calculations. By hard experience, theygain a better sense of whether their partners will adhere to bargains, evenwhen those bargains are inconvenient, or drop them at an opportune moment.Moreover, the density of their mutual ties gives them powerful levers to ensurecompliance. Close, multiplex ties may even promote shared expectations andbeliefs, which, in turn, strengthen the bargains themselves in a virtuous circle.

This cooperative framework, together with the protection of property andcontractual rights in liberal democracies, profoundly affects the global organi-zation of production and trade. States and corporations can seek out the largergains that come from specialization and an international division of labor.

T H E A R G U M E N T I N A N U T S H E L L 13

These gains are secured within a web of voluntary, self-enforcing cooperation,at least among established democracies.

These are not just gains from economic exchange. They are gains from ex-changing all kinds of commitments, from trade to environmental issues, fromnational security to criminal justice. In all these areas and many more, solidagreements encourage democracies to depend on each other in the widestsense. They also provide a more stable framework for private citizens andfirms to travel, trade, provide services, and invest abroad.

This intermeshing of civil societies is an important by-product of the demo-cratic peace and doubtless reinforces it. The development of global corpora-tions, for example, makes domestic politics more visible to outsiders and raisesthe costs of breaking economic ties. But forming such links in the first placerequires an overarching political framework based on interstate arrangements,a framework that establishes rules for commerce as well as institutions fordispute resolution.

To capture these gains from specialization and interdependence, states mustnot only make agreements, they must be able to rely on them. Otherwise, theagreements would be hollow, little more than empty promises to ensnare thenaıve and trap the ill informed. Genuine gains from trade come only whenagreements can be counted upon with some confidence. It is this reliance thatinduces real changes in actors’ behavior and a more efficient allocation ofresources.

The problem, of course, is that it is often risky to rely on internationalagreements. To rely on others is to become more vulnerable to their choices,including the chance that some partners will simply break their bargains whenit suits them. Since there are no arrangements to compensate innocent parties(as there usually are in domestic contracts), states must take care or suffer theconsequences.

States constantly struggle to protect themselves from these risks while stillcapturing the benefits from international cooperation. They do both simultane-ously, and our theories need to recognize this duality. In a world where statesare interconnected strategically and economically, they have twin concerns:self-protection and self-improvement. Self-improvement, mainly greater eco-nomic prosperity, is valued in its own right by modern states. Elected leadersare held accountable for its successful pursuit, just as they are held accountablefor national security. In a world where technological advances are central tomilitary effectiveness, prosperity can be a source of security. But this does notmean economic growth is merely the handmaiden of military effectiveness oris sought mainly for that purpose. It is equally true that states value self-protec-tion so that they can pursue greater wealth and freedom, and other aspects ofself-improvement.

Sometimes self-improvement is purely a domestic affair, but often it is not.It requires some measure of international cooperation and coordination, which

C H A P T E R 114

supports a wide range of profitable exchanges. Even under the most difficultcircumstances, states try to figure out ways to capture these gains from ex-change if only they can adequately contain the risks.

These risks never completely vanish, no matter what kinds of states areinvolved. On the other hand, the risks are not constant either. They are signifi-cantly lower when well-established constitutional democracies deal with eachother. The reason—to reiterate the main argument of the book—is that thesestates are more reliable partners because of four basic features of their statestructures:

1. high transparency, which allows outsiders as well as interested citizensto observe policy choices, grand strategies, and major regime discontinuitiesin a timely way, as well as to see the sources and intensity of support andopposition to specific commitments;

3. continuity of governance, based on clear rules for selecting senior of-ficials and ensuring orderly succession, so policies evolve within the frame-work of stable domestic regimes;

3. high audience costs, which political leaders must pay if they breakimportant promises to a mass electorate; and

4. constitutional governance, in which settled arrangements define andlimit the powers of public officials and ensure due process in official acts,including government commitments to its own citizens and to other states.

These complementary attributes raise the confidence of partners and reducethe dangers of surprise defections.27 They are basic—and unique—elements ofconstitutional democratic governance. Taken together, they form the con-tracting advantage of democracies. Democracies do have disagreements andconflicts of interest, sometimes serious ones. But they are also more reliablepartners, better equipped to settle disputes among themselves by durableagreements short of war.

These advantages cumulate in two ways. First, they are reflected in multiplebargains and extensive relationships among democracies, especially in eco-nomic and security affairs. These bargains not only provide them with moreinformation about each other; they give them multiple tools to ensure compli-ance. Their individual bargains do not stand alone. They are embedded in adense network of profitable relationships. Democracies are not bound togetherby one thick rope but by Velcro, with its innumerable strands. Second, as theserelationships develop over time, partners learn more about each other. If theirrelations are successful and survive temptations to defect for short-term gains,they build mutual confidence. These multiple bargains, and democracies’ ca-pacity to learn from them, give them unique capacities to manage relationships,avoid the costly escalation of disputes, and settle even serious conflicts withother democracies by durable agreements rather than by deadly fighting.

T H E A R G U M E N T I N A N U T S H E L L 15

Figure 1.1 Contracting Basis of the Democratic Peace

This explanation is summarized in figure 1.1.

A Roadmap for the Book

That is the argument in a nutshell. The remainder of the book explains thecausal mechanisms, evaluates them in light of the empirical evidence, andconsiders alternative explanations. After that, I extend the argument by show-ing how secure contracting affects numerous relationships among democraciesbeyond war and peace. Before delving into all this, however, we need to besure we are explaining something real and important.

In the chapter that follows, I ask, “Is There Really Peace among Democra-cies?” Some scholars have protested that peace among democracies is an illu-sion, that the findings are statistically insignificant. Before trying to explainthe democratic peace, I need to show that it is real and not a statistical quirk.In chapter 2, I briefly review the empirical evidence and consider its mostimportant critiques. I reach two main conclusions. First, democracies really dofight each other less often. A wide range of empirical studies show that thereis something real here to be explained. Second, existing explanations do notadequately capture the political logic of the democratic peace. In many cases,democracies have threatened each other with force and sometimes only nar-rowly averted war. Critics have examined some of these “close calls” andshown that the diplomacy looked very different from benign visions of thedemocratic peace. From that, they infer that the democratic peace is a mirage,that it is just Realpolitik plus dumb luck. That is wrong, I think, because itcannot explain why war between democracies is systematically less frequent.Still, the critics do raise a legitimate challenge. A convincing explanation notonly needs to predict the outcomes accurately, it needs to explain the causesin a way that tracks and clarifies the diplomatic record and makes sense out ofstates’ actions.

C H A P T E R 116

Chapter 3, “A Contracting Theory of the Democratic Peace,” outlines myown explanation, beginning with a rationalist theory of war. It shows why,according to such a theory, constitutional democracies should fight each otherless often. The essence of the rationalist theory is that war stems from private(hidden) information and commitment problems, the inability to project prom-ises dependably into the future.28 Stable constitutional democracies have sys-tematic advantages in both areas, which should allow them to avoid war byreliable bargains. Getting to these bargains, however, may not be an easy ride.Although democracies have advantages, they still have some hidden informa-tion, some lingering doubts about compliance, and a normal desire to win thelargest share of any joint gains. That means democracies will bargain hard and,in the process, could well threaten each other. On the other hand, they arebetter equipped than other states to settle these disputes without incurring thegrave costs and risks of war.

In chapters 4 and 5, I analyze the sources of this contracting advantage ingreater depth and discuss the general problem of self-enforcing internationalbargains. These chapters develop the causal logic and show how democraciesuse their contracting advantages to forge dependable partnerships. This expla-nation incorporates some features of existing accounts of the democratic peace,which stress either democratic values or institutions, and folds them into amore comprehensive explanation. Chapter 4 is entitled “Why Democratic Bar-gains Are Reliable: Constitutions, Open Politics, and the Electorate.” Chapter5 is “Leadership Succession as a Cause of War.”

This explanation, as I have already mentioned, predicts that democracieswill have other distinctive relationships beyond the fundamental achievementof peace. The contracting advantages of democracies should yield a web ofagreements, covering a wide range of mutually profitable exchanges. Chapter6 evaluates these predictions as well as the limits of the democratic peace.

Chapter 7 is the conclusion, “Reliable Partners and Reliable Peace.” Itquickly reviews the findings of the book, presents them in a straightforwardtable, and discusses the implications for world politics and for policymakersin democratic states.

Our first task, however, is to make sure we are explaining something real.In the next chapter we ask: Is the democratic peace a meaningful finding or astatistical fluke?